I personally think coders get better with age, like lounge singers.
Nailed it :)
“Peter Steinberger is a great example of how AI is catnip very specifically for middle-aged tech guys. they spend their 20s and 30s writing code, burn out or do management stuff for a decade, then come back in their late 40s/50s and want to try to throw that fastball again. Claude Code makes them feel like they still got it.”
This describes me nearly perfectly. Though I didn’t exactly burn out of coding, I accidentally stumbled upon being an EM while I was coding well and enjoying. But being EM stuck so I got into managing team(s) at biggish companies which means doing everything except one that I enjoy the most which is coding.
However now that I run my own startup I’m back to enjoying coding immensely because Claude takes care of grunt work of writing code while allowing me to focus on architecture, orchestration etc. Immense fun.
My experience is that people who weren't very good at writing software are the ones now "most excited" to "create" with a LLM.
I've been a tech lead for years and have written business critical code many times. I don't ever want to go back to writing code. I am feeling supremely empowered to go 100x faster. My contribution is still judgement, taste, architecture, etc. And the models will keep getting better. And as a result, I'll want to (and be able to) do even more.
I also absolutely LOVE that non-programmers have access to this stuff now too. I am always in favor of tools that democratize abilities.
Any "idiot" can build their own software tailored to how their brains think, without having to assemble gobs of money to hire expensive software people. Most of them were never going to hire a programmer anyway. Those ideas would've died in their heads.
One thing is for sure LLMs will bring down down the cost of software per some unit and increase the volume.
But..cost = revenue. What is a cost to one party is a revenue to another party. The revenue is what pays salaries.
So when software costs go down the revenues will go down too. When revenues go down lay offs will happen, salary cuts will happen.
This is not fictional. Markets already reacted to this and many software service companies took a hit.
What's missing from this is that iconic phrase that all the AI fans love to use: "I'm just having fun!"
This AI craze reminds me of a friend. He was always artistic but because of the way life goes he never really had opportunity to actively pursue art and drawing skills. When AI first came out, and specifically MidJourney he was super excited about it, used it a lot to make tons and tons of pictures for everything that his mind could think of. However, after awhile this excitement waned and he realized that he didn't actually learn anything at all. At that point he decided to find some time and spend more time practicing drawing to be able to make things by himself with his own skills, not by some chip on the other side of the world and he greatly improved in the past couple of years.
So, AI can certainly help create all the "fun!!!" projects for people who just want to see the end result, but in the end would they actually learn anything?
I'm having a lot of fun with AI. Any idiot can't prompt their way to the same software I can write. Not yet anyways.
I'm excited to work with AI. Why? Because it magnifies the thing I do well: Make technical decisions. Coding is ONE place I do that, but architecture, debugging etc. All use that same skill. Making good technical decisions.
And if you can make good choices, AI is a MEGA force multiplier. You just have to be willing to let go of the reins a hair.
I grew up without a mentor and my understanding of software stalled at certain points. When I couldn’t get a particular os API to work, in Google and stack overflow didn’t exist, and I had no one around me to ask. I wrote programs for years by just working around it.
After decades writing software I have done my best to be a mentor to those new to the field. My specialty is the ability to help people understand the technology they’re using, I’ve helped juniors understand and fix linker errors, engineers understand ARP poisoning, high school kids debug their robots. I’ve really enjoyed giving back.
But today, pretty much anyone except for a middle schooler could type their problems into a ChatGPT and get a more direct answer that I would be able to give. No one particularly needs mentorship as long as they know how to use an LLM correctly.
That said, I still feel strongly about mentorship though. It's just that you can spend your quality time with the busy person on higher-level things, like relationship building, rather than more basic questions.
In the hands of a knowledgeable engineer these tools can save a lot of drudge work because you have the experience to spot when they’re going off the rails.
Now imagine someone who doesn’t have the experience, and is not able to correct where necessary. Do you really think that’s going to end well?
Really?
The vibe coders are running into a dark forest with a bunch of lobsters (OpenClaw) getting lost and confused in their own tech debt and you're saying they can prompt their way to the same software?
Someone just ended up wiping their entire production database with Claude and you believe that your experience is for nothing, towards companies that need stable infrastructure and predictability.
Cognitive debt is a real thing and being unable to read / write code that is broken is going to be an increasing problem which experienced engineers can solve.
Do not fall for the AI agent hype.
I felt what you describe feeling. But it lasted like a week in December. Otherwise there’s still tons of stuff to build and my teams need me to design the systems and review their designs. And their prompt machine is not replacing my good sense. There’s plenty of engineering to do, even if the coding writes itself.
Staying up and re-learning what I used to love long ago has given me a new found passion as well. Even if I do vibe code some scripts, at least I have the background now to go through them and make sure they make sense. They're things I'm using in my own homelab and not something that I'm trying to spin up a Github repo for. I'm not shipping anything. I'm refreshing my old skills and trying to bring some of them up to date. An unfortunate reality is that my healthcare career is going to be limited due to multiple injuries along the way, and I need to try to be as current as I can in case something happens. My safety net is limited.
I’m probably going to go back and redo everything with my own code.
1. Creating something
2. Solving puzzles
3. Learning new things
If you are primarily motivated by seeing a finished product of some sort, then I think agentic coding is transcendent. You can get an output so much quicker.
If your enjoyment comes from solving hard puzzles, digging into algorithms, how hardware works, weird machine quirks, language internals etc... then you're going to lose nearly all of that fun.
And learning new things is somewhere in the middle. I do think that you can use agentic coding to learn new technologies. I have found llms to be a phenomenal tool for teaching me things, exploring new concepts, and showing me where to go to read more from human authors. But I have to concede that the best way to learn is by doing so you will probably lose out on some depth and stickiness if you're not the one implementing something in a new technology.
Of course most people find joy in some mix of all three. And exactly what they're looking for might change from project to project. I'm curious if you were leaning more towards 2 and 3 in your recent project and that's why you were so unsatisfied with Claude Code.
I guess if you're in an iterative MVP mindset then this matters less, but that model has always made me a little queasy. I like testing and verifying the crap out of my stuff so that when I hand it off I know it's the best effort I could possibly give.
Relying on AI code denies me the deep knowledge I need to feel that level of pride and confidence. And if I'm going to take the time to read, test and verify the AI code to that level, then I might as well write most of it unless it's really repetitive.
It's a different conversation when we talk about people learning to code now though. I'd probably not recommend going for the power tool until you have a solid understanding of the manual tools.
But then it makes me ask if the agents will get so good that craftsmanship is a given? Then that concern goes away. When I use Go I don't worry too much about craftsmanship of the language because it was written by a lot of smart people and has proven itself to be good in production for thousands of orgs. Is there a point at which agents prove themselves capable enough that we start trusting in their craftsmanship? There's a long way to go, but I don't think that's impossible.
Like just yesterday I started to notice the increasing pressure of an increasingly hard-to-navigate number of Claude chats. So I went searching for something to organize them. I did find an extension, but it's for Chrome, and I'm a Firefox person, so I had Claude look at it with the initial idea of porting to Firefox. Then in the analysis, Claude mentioned creating an extension from scratch, and that's what I went for.
I've never really used JavaScript, let alone created a Firefox extension before, but in a few minutes I was iterating on one, figuring out how I wanted it to work with Claude, and now I have a very nice and featureful chats organizer. And I haven't even peeked at the code. I also now have a firm idea of this general spec of how I want arbitrary list-organizing UI to look+behave going forward.
I think of AI like a microdose of Speed Force. Having super speed doesn't mean you don't like running; it just means you can run further and more often. That in turn justifies a greater amount of time spent running.
Without the Speed Force, most of the time you were reliant on vehicles (i.e. paying for third-party solutions) to get where you needed to go. With the Speed Force, not only can you suddenly meet a lot more of your transportation needs by foot, you're able to run to entirely new destinations that you'd never before considered. Eventually, you may find yourself planning trips to yet unexplored faraway harsh terrains.
If your joy in running came from attempting to push your biological physical limits, maybe you hate the Speed Force. If you enjoy spending time running and navigating unfamiliar territory, the Speed Force can give you more of that.
Sure, there are also oddballs who don't know how to run, yet insist on using the Speed Force to awkwardly jump somewhere vaguely in the vicinity of their destination. No one's saying they don't exist, but that's a completely different crowd from experienced speedsters.
As someone who enjoys technology, and using it, and can just barely sort-of code but really not, agentic coding must be wonderful. I have barely scratched the surface with a couple of scripts. But simply translating "here's what I want, and how I would have done it the last time I used Linux 20 years ago, show me how to do it with systemd" is so much easier than digging through years of forum posts and trying to make sure they haven't all been obsoleted.
None of it is new. None of it is fancy. I do regret that people aren't getting credit for their work, but "automount this SMB share from my NAS" isn't going to make anyone's reputation. It's just going to make my day easier. I really did learn enough to set up a NAT system to share a DSL connection with an office in the late 1990s on OpenBSD. It took a long time, and I don't have that kind of free time anymore. I will never git gud. It's this, or just be another luser who goes without.
The "creating something" idea... That's more complex. With agentic coding something can be created, but did I create it? Using agentic coding feels like hiring someone to do the work for me. For example, I just had all the windows in my house replaced. A crew came out at did it. The job is done, but I didn't do anything and felt no pride or sense of accomplishment in having these new windows. It just happened. Contrast that to a slow drain I had in my bathroom. I took the pipes apart, found the blockage, cleared it out, and reassembled the drain. When I next used the sink and the water effortlessly flowed away, I felt like I accomplished something, because I did it, not some plumber I hired.
So it isn't even about learning or solving puzzles, it's about being the person who actually did the work and seeing the result of that effort.
The inherent value of creating is something I was missing. Solving puzzles might be part of that, but not all. It's the classic Platonic question about how we value actions: for their own sake, for their results, or for both.
I think we agree that coding can be both, and it sounds like you feel the value for its own sake is lackluster in agentic coding -- It's just too easy. And I think that's the core sliding scale: Do you value creation more for its own sake or for its results? Where you land on that spectrum probably influences how people feel about agentic coding.
That being said, I also think that agentic coding can give enough of a challenge to scratch the itch of intrinsic value of creating. To a certain degree I think it's about moving up the abstraction chain to work more on architecture and product design. Those things can be fun and rewarding too. But fundamentally it's a preference.
I did put in 2 days of work to come up with what Claude used to ultimately do what it did... but when I look at the resulting code, I feel nothing. Having the idea isn't the same as being the one who actually did the thing. I plan to delete the branch next week. I don't want to maintain what it did, and think it should be less complex than it made it.
I will add though, on 2 and 3, during most of the coding I do in my day job as a staff engineer, it’s pretty rare for me to encounter deeply interesting puzzles and really interesting things to learn. It’s not like I’m writing a compiler or and OS kernel or something; this is web dev and infra at a mid sized company. For 95% of coding tasks I do I’ve seen some variation already before and they are boring. It’s nice to have Claude power through them.
On system design and architecture, the problems still tend to be a bit more novel. I still learn things there. Claude is helpful, but not as helpful as it is for the code.
I do get the sense that some folks enjoy solving variations of familiar programming puzzles over and over again, and Claude kills that for them. That’s not me at all. I like novelty and I hate solving the same thing twice. Different tastes, I guess.
"If your identity is tied to you being an iOS developer, you are going to have a rough time. But if your identity is 'I'm a builder!' it is a very exciting time to be alive."
Plus, there is no rule that says you can't keep coding if it's faster for you and/or it's quicker in general. e.g I can write a Perl one liner much faster than Claude can. Heck, even if it's not faster and you enjoy coding, just keep coding.
Also, when I write code myself, I still ask Claude to review it. It's faster than asking a human colleague to review it, so you can have Claude review often. Just today after a five-minute review Claude said a piece of code I wrote had four bugs, three of which were hallucinations and one was a real bug. I honestly do think it would have taken me a bit more than five minutes to find that one real bug.
Over the past couple months, I've created several applications with Claude Code. Personal projects that would've taken me weeks, months, or possibly forever, since I generally get distracted and move on to something else. I write pretty decent specs, break things into phases, and make sure each phase is solid before moving on to the next.
I have Claude build things in frameworks I would've never tried myself, just because it can. I do actually look at the code. Some of it is slop. In a few cases, it looks like it works, but it'll be a totally naive or insecure implementation. If I really don't like how it did something, I'll revert and give it another attempt. I also have other AIs review it and make suggestions.
It's fun, but I ultimately gain little intellectual satisfaction from it. It's not like the old days at all. I don't feel like I'm growing my skill set. Yes, I learned "something", but it's more about the capabilities of AI, not the end result.
Still, I'm convinced this is the future. Experienced developers are in the best position to work with AI. We also may not have a choice.
How do you think your company's CEO is going to feel when you tell them you could be finishing the software much faster, but you'd rather not, because it feels better to do it by hand?
For work, companies won't support it. Get it done. Fast. That's the new norm.
There should also be a symbiotic relationship at a job. Yes, they get something from me, but I should also get something… learning and some amount of satisfaction… in addition to the paycheck. I can get a paycheck anywhere.
It’s not the “new norm” unless employees accept it as the new normal. I don’t know why anyone would accept a completely one-sided situation like that.
How do you function on a team, where you have to maintain code others have written?
There are only 3 or 4 of us working on most of the code I touch. 3 of us have worked together in some form or another for close to 20 years.
So excited to be getting to my backlog of apps that I've wanted but couldn't take the time to develop on my own. I'm 66 and have been in the software field in various capacities (but programming mostly as a hobby). Here's a partial list of apps I've completed in the last few months:
- Grocery List with some tracking of frequent purchases
- Health Log for medical history, doc appointments and past visits
- Habits Tracker with trends I’m interested
- Daily Wisdom Reader instead of having multiple ebooks to keep track of where I'm at
- A task manager similar to the old LifeBalance app
- A Home Inventory app so that I can track what I have, warranty, and maintenance
- An ios watch app to see when I'm asleep so that it can turn off my music or audiobook
- An ios watch chess tactics trainer app
- some games
Many of these are similar to paid offerings, but those didn't check off all the features I really wanted, so I vibe-coded my own. They all do what I want, the way I want it to.
Afaik there are no open source projects that do this. AWS has a behemoth of a distributed system you can deploy in order to do something similar. But I made a Python script that does it in an afternoon with a couple of prompts.
It's the kind of thing that would be hours of tedious work, then even more time to actually make all the changes to the account. Instead I just say "yeah do all of that" and it is done. Magic stuff. Thousands of lines of Python to hit the Amazon APIs that I've never even looked at.
I wouldn't trust thousands of lines of code from one of my co-workers without testing
And yes, I have occasionally run into compiler bugs in my career. That's one reason we test.
How did you verify that?
> prone to hallucination
You know humans can hallucinate?
> is perfectly deterministic
We agree then that you can verify, test, and trust the deterministic code an LLM produces without ever looking at it.
> That's one reason we test
That's one way we can trust and verify code produced by an LLM. You can't stop doing all the other things that aren't coding.
I get there's a difference. Shitty code can be produced by LLMs or humans. LLMs really can pump out the shitty code. I just think the argument that you cant trust code you haven't viewed is not a good argument. I very much trust a lot of code I've never seen, and yes I've been bitten by it too.
Not trying to be an ass, more trying to figure out how im going to deal for the next decade before retirement age. Uts going to be a lot of testing and verification I guess
- A "semantically enhanced" epub-to-markdown converter
- A web-based Markdown reader with integrated LLM reading guide generation (https://i.imgur.com/ledMTXw.png)
- A Zotero plugin for defining/clarifying selected words/sentences in context
- An epub-to-audiobook generator using Pocket TTS
- A Diddy Kong Racing model/texture extractor/viewer (https://i.imgur.com/jiTK8kI.png)
- A slimmed-down phpBB 2 "remake" in Bun.js/TypeScript
- An experimental SQLite extension for defining incremental materialized views
...And many more that are either too tiny, too idiosyncratic, or too day-job to name here. Some of these are one-off utilities, some are toys I'll never touch again, some are part of much bigger projects that I've been struggling to get any work done on, and so on.
I don't blame you for your cynicism, and I'm not blind to all of the criticism of LLMs and LLM code. I've had many times where I feel upset, skeptical, discouraged, and alienated because of these new developments. But also... it's a lot of fun and I can't stop coming up with ideas.
I think it's also somewhat addictive. I wonder if that's part of what's at play here.
A coworker that never argues with you, is happy to do endless toil... sometimes messes up but sometimes blows your mind...
* Implementing a raw Git reader is daunting.
* Codifying syntax highlighting rules is laborious.
* Developing a nice UI/UX is not super enjoyable for me.
* Hardening with latest security measures would be tricky.
* Crafting a templating language is time-consuming.
Being able to orchestrate and design the high-level architecture while letting the LLM take care of the details is extremely rewarding. Moving all my repositories away from GitLab, GitHub, and BitBucket to a single repo under my own control is priceless.
"in (language I'm familiar with) I use (some pattern or whatever) what's the equivalent in (other language)?"
It's really great for doing bits and then get it to explain or you look and see what's wrong and modify it and learn.
I can ask an LLM for specific help with my codebase and it can explain things in context and provide actual concrete relevant examples that make sense to me.
Then I can ask again for explanations about idiomatic code patterns that aren't familiar for me.
Working on my own, I don't get that feedback and code review loop.
Working with new languages and techniques, or diving into someone else's legacy code base is no longer as daunting with an LLM to ask for help!
Try to tell Claude Code to refactor some code and see if it doesn't just delete the entire file and rewrite it. Sure that's cute, but it's absolutely not okay in a real software environment.
I do find this stuff great for hobbyist projects. I don't know if I'd be willing to put money on the line yet
It's given me the guts to be a solo-founder (for now). I
There are definitely a lot of limitations with Claude Code, but it's fun to work through the issues, figure out Claude's behavior, and create guardrails and workarounds. I do think that a lot of the poor behavior that agents exhibit can be fixed with more guardrails and scaffolding... so I'm looking forward to the future.
Following this idea, what do people think "backend" work will involve? Building and tweaking models, and the infra around them? Obviously everyone will shift more into architecture and strategy, but in terms of hands-on technical work I'm interested in where people see this going.
"I used to write java code and the compiler turned it into JVM bytecode.
Now I write in English and the LLMs compile it into whatever language I want."
Although as one HN commenter pointed out: English is a pretty bad programming language as it's way more ambiguous than most programming languages.
Wake me when we have ethically trained, open source models that run locally. Preferably high-quality ones.
100% agree even with half your experience.
But, uh, yeah... I've been noticing a growing divide between people like OP who are either already retired or are wealthy enough that they could if they wanted to who absolutely love the new world of LLMs, and people who aren't currently financially secure and realize that LLMs are going to snatch their career away. Maybe not this year, but not too far out either.
How does the saying go again? "It takes a village to reach financially secure retirement"
Fucking wild.
I want a game that generates its own mechanics on the fly using AI. Generates itself live.
Infinite game with infinite content. Not like no mans sky where everything is painfully predictable and schematic to a fault. No. Something that generates a whole method of generating. Some kind of ultra flexible communication protocol between engine and AI generator that is trained to program that protocol.
Develop it into a framework.
Use that framework to create one game. A dwarf fortress adventure mode 2.0
I have no other desires, I have no other goals, I don’t care. I or better yet - someone else, must do it.
If the software produced is for internal use, the point is probably moot. But if it isn't, this seems like a question that needs to be answered ASAP.
Claude Code and it's parallels have extinguished multiple ones.
I was able to steer clear of the Bitcoin/NFT/Passport bros but it turns out they infiltrated the profession and their starry puppy delusional eyes are trying to tell me that iteration X of product Y released yesterday evening is "going to change everything".
They have started redefining what "I have build this" actually means, and they have outjerked the executives by slinging outrageous value creation narratives.
> I’m chasing the midnight hour and not getting any sleep.
You are 60; go spend some time with your grand-kids, smell a flower, touch grass forget chasing anything at this age cause a Tuesday like the others things are gonna wrap up.
Absolutely sincerely.
